/hibernated

HibernateD is ORM for D language (similar to Hibernate)

Primary LanguageD

HibernateD

DUB Package GitHub CI Gitter

HibernateD is ORM for D language (similar to Hibernate)

The project is hosted on github with documentation available on the wiki.

Uses DDBC as an abstraction layer for the underlying database. DDBC currently supports SQLite, MySQL, Postgres, SQL Server, and potentially Oracle.

SQL dialects supported by Hibernated

Currently hiberanted supports:

  • SQLite (requires SQLite v3.7.11 or later)
  • MySQL (tests are run against mysql v5.7)
  • Postgres

But since the underlying DDBC library now has ODBC support some work is underway to also add

  • SQL Server (2017 & 2019)
  • Oracle (due to the size of the docker image for Oracle this may not be tested during CI)

Sample code:

import hibernated.core;
import std.algorithm;


// Annotations of entity classes

class User {
    long id;
    string name;
    Customer customer;
    @ManyToMany // cannot be inferred, requires annotation
        LazyCollection!Role roles;
}

class Customer {
    int id;
    string name;
    // Embedded is inferred from type of Address
    Address address;

    Lazy!AccountType accountType; // ManyToOne inferred

    User[] users; // OneToMany inferred

    this() {
        address = new Address();
    }
}

@Embeddable
class Address {
    string zip;
    string city;
    string streetAddress;
}

class AccountType {
    int id;
    string name;
}

class Role {
    int id;
    string name;
    @ManyToMany // w/o this annotation will be OneToMany by convention
        LazyCollection!User users;
}

int main() {

    // create metadata from annotations
    EntityMetaData schema = new SchemaInfoImpl!(User, Customer, AccountType, 
            Address, Role);




    // setup DB connection factory
    version (USE_MYSQL) {
        import ddbc.drivers.mysqlddbc;
        MySQLDriver driver = new MySQLDriver();
        string url = MySQLDriver.generateUrl("localhost", 3306, "test_db");
        string[string] params = MySQLDriver.setUserAndPassword("testuser", "testpasswd");
        Dialect dialect = new MySQLDialect();
    } else {
        import ddbc.drivers.sqliteddbc;
        SQLITEDriver driver = new SQLITEDriver();
        string url = "zzz.db"; // file with DB
        static import std.file;
        if (std.file.exists(url))
            std.file.remove(url); // remove old DB file
        string[string] params;
        Dialect dialect = new SQLiteDialect();
    }
    DataSource ds = new ConnectionPoolDataSourceImpl(driver, url, params);


    // create session factory
    SessionFactory factory = new SessionFactoryImpl(schema, dialect, ds);
    scope(exit) factory.close();

    // Create schema if necessary
    {
        // get connection
        Connection conn = ds.getConnection();
        scope(exit) conn.close();
        // create tables if not exist
        factory.getDBMetaData().updateDBSchema(conn, false, true);
    }

    // Now you can use HibernateD

    // create session
    Session sess = factory.openSession();
    scope(exit) sess.close();

    // use session to access DB

    // read all users using query
    Query q = sess.createQuery("FROM User ORDER BY name");
    User[] list = q.list!User();

    // create sample data
    Role r10 = new Role();
    r10.name = "role10";
    Role r11 = new Role();
    r11.name = "role11";
    Customer c10 = new Customer();
    c10.name = "Customer 10";
    c10.address = new Address();
    c10.address.zip = "12345";
    c10.address.city = "New York";
    c10.address.streetAddress = "Baker st., 12";
    User u10 = new User();
    u10.name = "Alex";
    u10.customer = c10;
    u10.roles = [r10, r11];
    sess.save(r10);
    sess.save(r11);
    sess.save(c10);
    sess.save(u10);

    // load and check data
    User u11 = sess.createQuery("FROM User WHERE name=:Name").
        setParameter("Name", "Alex").uniqueResult!User();
    assert(u11.roles.length == 2);
    assert(u11.roles[0].name == "role10" || u11.roles.get()[0].name == "role11");
    assert(u11.roles[1].name == "role10" || u11.roles.get()[1].name == "role11");
    assert(u11.customer.name == "Customer 10");
    assert(u11.customer.users.length == 1);
    assert(u11.customer.users[0] == u10);
    assert(u11.roles[0].users.length == 1);
    assert(u11.roles[0].users[0] == u10);

    // remove reference
    u11.roles = u11.roles().remove(0);
    sess.update(u11);

    // remove entity
    sess.remove(u11);
    return 0;
}

Additional Features

Composite Keys

If a database contains tables with a composite primary key, the @EmbeddedId can be used to represent this. The columns that represent the primary key should be in an @Embeddable class which is then referenced in a property annotated with @EmbeddedId.

For example, consider a database table created via the following SQL command:

CREATE TABLE invoices (
    vendor_no VARCHAR(8) NOT NULL,
    invoice_no VARCHAR(20) NOT NULL,
    amount_e4 INTEGER);
ALTER TABLE invoices
    ADD CONSTRAINT invoices_pkey PRIMARY KEY (vendor_no, invoice_no);

To represent this using HibernateD, the following code would be used:

@Embeddable
class InvoiceId {
    string vendorNo;
    string invoiceNo;

    // Be sure to implement this to benefit from session caching.
    bool opEquals(const InvoiceId o) const @safe {
        return vendorNo == o.vendorNo && invoiceNo == o.invoiceNo;
    }
}

@Table("invoices")
class Invoice {
    @EmbeddedId InvoiceId invoiceId;
    int amountE4;
}

Note: At the time of writing, there are important limitations.

  1. The function DBInfo.updateDbSchema(Connection conn, bool dropTables, bool createTables) does not generate schemas with compound keys.
  2. The Hibernate annotation @JoinColumns (plural) has not yet been implemented, thus, the @ManyToOne and @ManyToMany relations are not usable for classes using an @EmbeddedId.
  3. The @Embedded class referenced via an @EmbeddedId property should implement opEquals in order to gain performance benefits from session caching. These features will be added in future updates.